OCR Output

b. Image Analysis

An important aspect in the interpretation of multispectral image sets relates to how these are
acguired. Commercially available cameras are designed to provide aesthetically pleasing
images and not the scientific analysis of artworks, and as a result can often introduce
unwanted modifications to the multispectral images acquired. These include built-in
adjustments to the contrast, brightness, sharpness, gain controls and white balance, among
other factors, which modify the image taken by the camera into one closer to how the eye
perceives it. However, these amendments may cause the erroneous interpretation of the
information contained within multispectral images. This is particularly true for luminescence
images, as described below. If uncontrolled, these changes can have a substantial impact
on the reliability and reproducibility of images, making effective and meaningful comparisons
between images taken at different times and with different set-ups challenging, if not
impossible.

Figure 1-4 shows a detail from UV-luminescence images of a fresco by Agnolo Gaddi taken
at the basilica di Santa Croce in Florence by two different institutions, the Opificio delle
Pietre Dure (OPD) and the Getty Conservation Institute (GCI). The differences between
these images, which depict the same subject are clear and underline the need to address
the issues discussed.

Figure 1-4. UV-induced luminescence images of a detail from a fresco by Agnolo Gaddi taken at the
basilica di Santa Croce in Florence by two different institutions, the Opificio delle Pietre Dure (©
Annette Keller at the OPD, left) and the Getty Conservation Institute (© Organic Material in Wall
Painting Project, Getty Conservation Institute, right).

Version No. 1.0 Date : 14/10/2013

NTH FRAMEWORK